Bug 493025. Don't assert if we happen to do bitblit scroll analysis on some frames that are dirty because a reflow got interrupted. r+sr=roc
authorBoris Zbarsky <bzbarsky@mit.edu>
Thu, 14 May 2009 21:40:29 -0400
changeset 28359 42a2fc83189e
parent 28358 5f4ee25d609c
child 28360 39136779dd1d
push id7027
push userbzbarsky@mozilla.com
push date2009-05-15 01:41 +0000
treeherdermozilla-central@42a2fc83189e [default view] [failures only]
perfherder[talos] [build metrics] [platform microbench] (compared to previous push)
bugs493025
milestone1.9.2a1pre
Bug 493025. Don't assert if we happen to do bitblit scroll analysis on some frames that are dirty because a reflow got interrupted. r+sr=roc
layout/base/nsLayoutUtils.cpp
--- a/layout/base/nsLayoutUtils.cpp
+++ b/layout/base/nsLayoutUtils.cpp
@@ -1212,16 +1212,18 @@ nsLayoutUtils::ComputeRepaintRegionForCo
                                            nsIFrame* aMovingFrame,
                                            nsPoint aDelta,
                                            const nsRect& aCopyRect,
                                            nsRegion* aRepaintRegion)
 {
   NS_ASSERTION(aRootFrame != aMovingFrame,
                "The root frame shouldn't be the one that's moving, that makes no sense");
 
+  nsAutoDisableGetUsedXAssertions disableAssert;
+
   // Build the 'after' display list over the whole area of interest.
   // (We have to build the 'after' display list because the frame/view
   // hierarchy has already been updated for the move.
   // We need to ensure that the non-moving frame display items we get
   // are the same ones we would have gotten if we had constructed the
   // 'before' display list. So opaque moving items are only considered to
   // cover the intersection of their old and new bounds (see
   // nsDisplayItem::OptimizeVisibility). A moving clip item is not allowed